From natural language to executable models. How ChatGPT writes a system dynamics model

Abstract

The process of creating a simulation model, in particular using the System Dynamics methodology, is a complex and time-intensive task. It begins with a clear understanding of the problem, followed by the identification of key variables, and culminates in the construction of a Stock and Flow Diagram. Traditionally, this process has been laborious, requiring significant manual effort to translate a conceptual understanding into a functional model. With the advent of advanced AI tools like ChatGPT and Collaborative AI, this process can be significantly streamlined. This paper presents a personalized ChatGPT prompt, DAVID© (Draw A Vensim Initial Draft) that can automate the creation of a complete Vensim model, from problem description to a fully functional simulation. The importance of this tool, which is easy to use and modify, lies in the fact that it allows people who are not experts in creating simulation models to create them without prior training or experience.